inactivity contributes to health maladies including breast and colon cancers directly, and numerous other cancers indirectly, in part by contributing to obesity. Ethnic minorities report higher rates of physical inactivity and obesity. Individually-focused physical activity (PA) interventions are initially effective. Up to 50% drop out within 6 months, and maintenance continues to decline after this time. Ecological models posit that environments do not support PA impede PA maintenance. There have been few investigations testing the combined effects of supportive environments with individually focused interventions that increase PA. The purpose of the proposed study is (1) to determine whether a 24 week social cohesion intervention (SOCO) is more effective for adopting walking for leisure exercise and transport in comparison to a delayed control (CONT), (2) to determine whether walking is more effectively maintained by participants who reside in highly supportive PA areas (HIGH) in comparison to those who reside in low supportive PA areas (LOW), and (3) to determine whether this effect is transculturally replicable across 2 cultural groups from 2 cities. African American women in Houston and Hispanic women in Austin will be recruited from churches and community centers and randomized to either SOCO or CONT. Participants'neighborhoods will be classified as HIGH or LOW supportive. Participants will complete written socio demographic, walking (International Physical Activity Questionnaire), and health measures at baseline (T1), post intervention (T2), and 6 month follow-up (T3). Participants will complete accelerometer assessments as a validity check. It is hypothesized that (1) participants in SOCO will report more walking at T2 than CONT, (2) SOCO participants residing in HIGH neighborhoods will report greater minutes of walking than LOW at T3, and (3) this effect will be transculturally replicable. If supported, hypotheses will suggest that the sustained effect of SOCO strategies is moderated by environmental supports.